PRELIMINARY DRAFT: NOT FOR CITATION OR QUOTATION WITHOUT PERMISSION

Contesting Accountability and Legitimacy in Non-State Regulatory Regimes

Julia Black1

Introduction

It has been said that governments can ‘puzzle as well as power’.2 Shamelessly

misappropriating this comment, it could be said that academic papers can do the same.

Some puzzle their way through an issue, raising more questions than they answer; others

‘power’ on through, setting out the path that others must follow to find a solution to

whatever problem is being addressed.

This paper is of the former type. The issue that it ponders is that of the accountability

and legitimacy of decentred regulatory regimes. Decentred regulatory regimes are those

in which the state is not the sole locus of authority, or indeed in which it plays no role at

all. They are marked by fragmentation, complexity and interdependence between actors,

in which state and non-state actors are both regulators and regulated, and their boundaries

are marked by the issues or problems which they are concerned with, rather than

necessarily by a common solution. Such regimes pose a number of challenges which

writers across a range of disciplines - law, political science, international relations,

development studies - are all engaged in delineating and addressing. Indeed the issues to

which the ‘governance turn’ is giving rise is drawing commentators like moths round a

light.

1 Law Department and Research Associate, ESRC Centre for the Analysis of Risk and Regulation, London School of Economics and Political Science. This is an extremely preliminary draft and so is not for citation or quotation without permission.2 Heclo, H. (1974), Modern Social Politics in Britain and Sweden, New Haven: Yale University Press.

These challenges include the functional, the systemic, the democractic and the normative.

Functional challenges revolve around the problem of coordination; networks of

organisations within a regulatory regime may be characterised by complex

interdependencies and may lack a central locus of authority. There may not be a body

whose role it is to act as the lead ‘interpreter’ of the regimes’ rules or principles, for

example, or to otherwise steer or coordinate the activities of the multiple participants in

such a way that the regime moves towards the resolution of the problem which it both

defines and is defined by. Systemic challenges revolve around issues of fragmentation of

social systems, and the challenges that decentred regulation can pose to particular social

systems. For lawyers this is particularly the challenge posed to the identification and

identify of law by the presence of numerous normative orders, an issue debated in

international law in terms of the nature of ‘soft law’ and in legal theory journals in terms

of the challenges of legal pluralism. Democratic challenges arise from issues of

representation: who should be involved in the decision making structures of the various

components of the network; to whom should such bodies be accountable and how.

Normative challenges stems from normative concerns as to the goals and operation of the

regulatory regime: what conception of ‘the good’ is and should be pursued. Of these, the

functional, democratic and normative challenges in particular are often articulated in

terms of the accountability and legitimacy of the regimes as a whole and the different

actors within them.

For many engaged in that debate, the solution is to find functional equivalents to the

structures of accountability which are to be found in constitutional settlements, at least of

liberal democratic states. Checks and balances, dispute resolution processes,

democratisation – all are common features of the ‘powering’ solutions being advocated

by commentators in a range of disciplines.

This paper takes a different perspective. The usual quartet of accountability questions:

how, with respect to what, to whom and when, are not addressed, at least not directly.

The paper asks not what mechanisms are necessary to make actors in decentred

regulatory regimes accountable or even legitimate, nor even in accordance with what

values, or with respect to whom should they be made accountable in the regulatory

process. Rather it takes the perspective of those on whom the demands are being made,

and asks how they respond. What happens when these different accountability and

legitimacy demands are made, and indeed what role to the objects of the accountability

and legitimacy demands play in shaping those demands? This is of interest in its own

right; however it is also the logically prior question that has to be asked before any

‘powering’ proposals can be made.

The ‘puzzling’ proceeds in four main stages, and has two key consequences, from which

a number of further implications flow. First, it looks at the relationship between

accountability and legitimacy, and suggests that both are relational concepts which are

socially and discursively constituted. Secondly, it suggests in turn that actors within and

outside the regulatory regime have different perceptions as to the relevance and validity

of different legitimacy claims with respect to different regulatory actors, in other words

that there are different legitimacy communities. Thirdly, that different legitimacy claims,

and associated discourses, are not always compatible but may compete. Fourthly, that

although organisations can often participate in a number of different legitimacy

discourses simultaneously, and thus satisfy a range of different legitimacy communities,

this can not only have a deleterious affect on the organisation (which may suffer

‘multiple accountability disorder’) the differences between communities may be such that

organisations can face a legitimacy dilemma: that actions that they need to take to render

them legitimate for one legitimacy community are in direct opposition from those they

need to adopt to satisfy another.

So far, so familiar, perhaps. However, the paper then seeks to explore two main

consequences of these propositions for regulatory accountability and legitimacy. These

are: first, that different accountability or legitimacy mechanisms are not necessarily

substitutable, as some may suggest, as not all will satisfy every legitimacy community.

Secondly, and more importantly for this paper, that how organisations respond to these

competing legitimacy demands is structured by the particular institutional context in

which the regulatory regime, and the individual organisation, operates. Regulators are

not ciphers – the insights of the ‘ungovernability’ of actors apply as much to them as

those they seek to regulate. They can be active participants in the debate on their own

accountability and legitimacy, not just a passive recipient. They may exhibit the same

strategies of avoidance, defiance, manipulation, compromise or acquiescence in response

to pressures for their accountability and legitimacy as any organisation does in response

to any norms which others seek to impose on them.3 This much is familiar from

institutional theory, but its insights are often separated from the more normative or

‘powering’ prescriptions for accountability.

The paper takes a further theoretical step and, pursuing the interplay of institutionalism

with discourse theory, suggests that in acquiescing in, or at least reaching a compromise

with, some legitimacy communities, the organisation indicates that it is prepared to

recognise those, whilst refusing to recognise others. This has implications for those who

are seeking to get the organisation to respond to their accountability and legitimacy

demands. Absent other key resources, the paper suggests that those legitimacy

communities whose demands have not been recognised will have to translate them into

the discourses of those that have been recognised if they are to build the relationships of

accountability with the organisation that they seek.

Finally, the paper draws on the preceding analysis to offer a challenging empirical

research agenda into the accountability and legitimacy of decentred regulatory regimes.

The broad parameters of the accountability and legitimacy debates

Before embarking on the puzzle, it is worth reminding ourselves of the significance and

wide ranging nature of the contemporary debate over the accountability and legitimacy of

organisations, including but not limited to non-state regulators. The accountability and

legitimacy of organisations, ranging from firms to national, international and

supranational governmental bodies, from charities to international non-governmental

3 See further below.

organisations (INGOs), from standard setting bodies to investment arbitrators, is the

subject of an intense debate. The ‘governance turn’ revealed a plethora of non-

governmental actors who were performing what had been traditionally seen as core

‘governmental’ functions: welfare provision and regulation, or who in much broader

terms are seen as exercising significant amounts of power over those both inside and

outside them. The demands for corporate social accountability, the calls for improving

the ‘representativeness’ or ‘transparency’ of international regulatory and standard setting

bodies, or for enhancing the ‘legitimacy’ of INGOs all have at their base the same central

concern: that power is being exercised in a way which is insufficiently accountable to

others. As a result, organisations are, to use Power’s evocative phrase, being turned

‘inside out’.4 The details of their internal decision making structures and processes,

including their incentive structures, audit and risk management processes, are seen as

critically relevant to those outside them.

These multiple demands for enhanced legitimacy and accountability can lead to a

coalescence of norms which have varying geographic and sectoral applicability. Firms,

particularly multi-nationals, are the subject of a number of codes relating to corporate

social responsibility and human rights, for example the UN Global Compact, the draft

United Nations Norms on the Responsibilities of Transnational Corporations and other

Business Enterprises, the OECD Guidelines for Multinational Enterprises and the ILO

Tripartite Declaration of Principles concerning Multinational Enterprises and Social

Policy. Firms are not alone. There have been increasing calls on INGOs to improve their

legitimacy and accountability. Partly in response INGOs have developed their own codes

of practice. Some of these are functional statements of best practice, but they include

provisions, albeit often in very general terms, which relate to accountability and

transparency. For example, the One World Trust Global Accountability Project identifies

four dimensions of accountability: transparency, participation, evaluation, and complaints

and response, detailing operational guidelines on how these can be translated into

practice.5 Other codes are far narrower. Specific codes exist for INGOs giving disaster

4 M. Power, The Risk Management of Everything (Demos, 2005).5 M. Blagescu, L. de la Casas and R. Lloyd, Pathways to Accountability: the GAP Framework (London: 2005), available from www.oneworldtrust.org.

relief, or responding to HIV/Aids,6 or in specific countries or regions.7 The above are

examples of self-imposed and self-managed codes. NGOs can also be subject of codes

formulated by governments, for example the EC Code of Conduct on Non Profit

Oganisations, which focuses on the potential use of non-profit organisations as vehicles

of terrorist finance or money laundering, although the range of its recommendations

belies this relatively narrow rationale: the registration, enhanced financial transparency,

accountability and oversight of all non-profit organisations.8

Amongst this plethora of actors, non-state regulators occupy a curious position. They

may be campaigning NGOs, and as such occupy a dual role as lobbyist and regulator, or

they may focus primarily on a broad ‘regulatory’ function, at least in the sense that they

set written norms for others to follow. They may have a clear organisational structure,

such as Transparency International, the Forest Stewardship Council, the standard setting

bodies such as the International Standards Organisation (ISO) or the transnational

financial regulatory organisations: IOSCO or BCBS, for example.9 Such structures may

be federated, such as Responsible Care, which has a central organisation but then allows

regional bodies to develop which shape the Code in ways which make it relevant for their

own regions.

In these cases there is at least an organisation which ‘owns’ the norms or codes that are

produced. However, it may be that there is no central organisational structure, rather

there is a body of written norms which firms themselves have decided to apply, but there

is no central locus of authority to which they can turn to discuss the proper interpretation

or application of the principles. The Equator Principles provide a good example. These

are a set of principles for sustainable development which many banks require the 6 Code of Conduct for the International Red Cross, the Red Crescent and NGOs in Disaster Relief; Code of Good Practice for Responding to HIV / Aids.7 For example the BOND Statement of Principles (BOND is the British Overseas NGOs for Development – a network of over 300 UK based organisations working in international development).8 Recommendation for Member States and a Framework for a Code of Conduct for Non-Profit Organisations to Enhance Transparency and Accountability in the Non Profit Sector to Prevent Terrorist Financing and other types of Financial Abuse, 2005; contained as an Annex in The Prevention and Fight Against Terrorist Financing through enhanced national level coordination and greater transparency of the non-profit sector, COM(2005) 620 final.9 IOSCO is the International Organisation of Securities Commissions; BCBS is the Basle Committee on Banking Standards, part of the Bank of International Settlements.

borrowers to comply with when issuing loans for infrastructure development, mainly in

the energy sector (dams, pipelines etc). There is no one organisation which is responsible

for issuing the Principles, interpreting or revising them, however. The ‘regulators’ are

the banks, regulating both themselves and others to ensure compliance with the

principles, at least in the initial loan documentation. Non-state regulatory regimes in this

case have the organisational form of co-ordinated systems of corporate social

responsibility, rather than taking the form of a single agency regulator which parallels a

governmental body. Although the ‘crisis’ of accountability of non-state regulators tends

to tar all with the same broad brush, different types of non-state regulatory regimes can

rather pose different challenges; in particular the absence of any single locus of authority

within a regulatory regime enhances the challenges of functionality and of accountability

and legitimacy, discussed further below.

Relationship between accountability and legitimacy

Nonetheless, as many have pointed out, relying on the traditional forms of accountability

either of state based regulators, or indeed of corporations, are simply inadequate for such

systems of regulation.10 In these debates, as indeed so far in this paper, the terms

accountability and legitimacy tend to be conflated. It is worth spending a brief time

exploring their relationship more fully, however, before moving on to examine the

different forms that demands for each can take.

Definitions of accountability abound, and there are almost as many definitions as there

are articles on the subject. As Mulgan has commented, it is a word which ‘now crops up

everywhere, performing all sorts of analytical and rhetorical tasks, and carrying most of

the burden of democratic governance’.11 Or as Bovens puts it: ‘As an icon, the concept

has become less useful for analytical purposes, and today resembles a garbage can filled

with good intentions, loosely defined concepts, and vague images of good governance.’12

10 Eg C. Scott, ‘Accountability in the Regulatory State’ (2000) 27(1) Jnl Law and Society 38.11 R. Mulgan, ‘Accountability: An Ever Expanding Concept?’ (2000) 78(3) Public Administration 555 at 555.12 M. Bovens, Analysing and Assessing Public Accountability: A Conceptual Framework’ (2006) European Governance Paper No C-06-01, p.5.

As Mulgan demonstrates, from its ‘core’ meaning of being called to account for one’s

actions through a social exchange by an external body or group which has authority over

one, it has expanded to include ‘internal’ accountability, to one’s own conscience or to

norms of professional ethics; ‘responsiveness’ by governments to demands made on them

by citizens; ‘control’ over institutions through the checks and balances of political

systems, and ‘dialogue’: public discussion between citizens.13 There is dispute over each

dimension: whether accountability is internal or purely external;14 whether it involves

interpersonal, social exchange or whether accountability can be to the impersonalised

operations of the market; whether it has to involve hierarchical relationships of authority,

and whether accountability means control, as Scott argues for example,15 or whether

something less, such as responsiveness or the ability to impose consequences on the

object of accountability suffices.16

Whilst it is not proposed to engage here in the debate on the ‘true’ meaning of

accountability, some notion of accountability has to be adopted, at least for the purposes

of this analysis. It is suggested that contrary to being seen as a form of control,

accountability should not be equated with control. Debates about whether a person is

‘internally’ accountable to their own moral sense, for example, are conflating

‘accountability’ with the constraints on action that are posed by institutional norms.

Similarly notions of being ‘accountable’ to the market refer rather to the organisation’s

responsiveness to the actions of market actors; not that it explains itself to the market, or

engages in any dialogic process at all.

Rather, a far narrower notion of accountability is suggested: to be accountable is to agree

to subject oneself to relationships of external scrutiny which can have consequences.17

Although his definition is arguably too prescriptive of the actual processes that should be

13 Ibid.14 M. Dubnick, ‘Clarifying Accountability: An Ethical Framework’ in C. Sampford and N. Preston, Public Sector Ethics (Routledge, 1998); A. Sinclair, ‘The Chameleon of Accountability’ (1995) 20 Accounting Organizations and Society 219.15 C. Scott, op.cit. 16 Eg Bovens, op. cit.17 See also R. Keohane, ‘Global Governance and Democratic Accountability’ in D. Held and M Koenig-Archibugi (eds), Taming Globalization (Oxford: Polity Press, 2003).

adopted, Boven’s definition expresses the kernel of accountability: ‘a relationship

between an actor and a forum, in which the actor has an obligation to explain and justify

his or her conduct, the forum can pose questions and pass judgement, and the actor may

face consequences.’ The emphasis on the forum is arguably too restrictive, and stems

perhaps from Bovens’ focus on public institutions; it does however highlight what it is

suggested is one of the central aspects of accountability: a relational responsiveness to

others.18

Accountability is usually in turn linked to another iconic word: legitimacy. Like

accountability, legitimacy is a ‘mother and apple pie’ concept which no one can argue

against. Also like accountability, there is a significant debate as to its meaning, but in

this case one which occupies legal and political theorists more than it does those

interested in public management. Again, a short cut through the debate is proposed. It is

suggested that by legitimacy is meant whether or not an institution or organisation is

perceived as having a ‘right to govern’ both by those it seeks to govern and those on

behalf it purports to govern.19 Legitimacy may be an objective fact, but it is socially

constructed.20 It rests on the acceptability and credibility of the organisation to those it

seeks to govern. Organisations may claim legitimacy, and may perform actions and enter

into relationships in order to gain it. But legitimacy is rooted in the acceptance of that

organisation by others.

Legitimacy is linked with authority. Here again a short cut is taken through a thicket of

political and legal theory. Relying on Raz, as developed by Coleman, by authority is

meant whether or not what an actor says or requires makes a ‘practical difference’ to the

way that others act or behave, and whether it does so simply by virtue of the actor saying

it. In other words, does the mere fact that a particular actor stipulates that a particular

course of conduct should be followed mean that others (though by no means all) will alter

their conduct not as a result of a rationalistic pursuit of preferences but principally out of

a sense of obligation, or because that actor is respected or esteemed within the 18 Eg Keohane, above. The notion has also entered the CSR debate, see for example M. Painter Morland, ‘Redefining Accountability as Relational Responsiveness’ (2006) 66 Journal of Business Ethics 89.19 See discussions in R. Barker, Political Legitimacy and the State (Oxford: OUP, 1990).20 W.R. Scott, Institutions and Organisations (2nd ed; Sage, 2001)

community for whatever reason.21 This does not imply that those subject to the legitimate

authority of someone issuing directives have an obligation to obey; but that the authority

is perceived as justified in issuing directives to them.22

Accountability is usually seen as a central element of legitimacy in the sense that to be

legitimate, an actor has to be accountable. However, this is too general a characterisation

of their relationship to be satisfactory. As discussed, legitimacy is a perceived right to

govern: how and why that perception is forged is another, highly complex issue

addressed in part below, but the reasons are grounded in social acceptance.

Accountability, for its part, is a particular type of relationship between different actors: in

which one gives account and another has the authority to impose consequences as a

result. In other words, that the response of the person to whom account is made will

make a ‘practical difference’ to the conduct of the one who gives account, either

retrospectively, prospectively or both. A more specific description of the relationship of

the two concepts, at least as they are defined above, is therefore that perceptions of the

right to govern (legitimacy) may depend (in whole or in part) on whether the actor is

accepted as having an appropriate accountability relationships with others, often

including, but not necessarily confined to, the person whose perception is in question.

Constructing accountability and legitimacy

The debate on the accountability and legitimacy of non-governmental regulators,

especially at the transnational level, often depicts organisational actors which are

somehow disconnected: autonomous bodies which are not grounded in any particular

sector of society or even in any one state. However, as institutional theory emphasises,

all actors are embedded in social institutions, from which in turn they derive their

21 The concept thus draws on the concept of authority developed in jurisprudence by Raz as elaborated by Coleman: J. Raz, The Authority of Law (Oxford: OUP, 1979); J. Coleman, ‘Incorporationism, Conventionality and the Practical Difference Thesis’ The Practical Difference Thesis (1998) 4 Legal Theory 381.22 See discussion in eg W. Sadursky, ‘Law’s Legitimacy and “Democracy Plus”’ (2006) 26(2) OJLS 377.

legitimacy. The conceptions of legitimacy and accountability outlined above have a

strong affinity with institutional theory, or at least some branches of it.

A minimalist definition of institutions is that institutions comprise cognitive and moral

structures, rules, norms, conventions or operating procedures which are regarded as

socially or legally binding but which are not self-enforcing.23 For political scientists

institutions also comprise the key political structures: legislature, executive, voting

system, legal system, and bureaucracy. For economists, they also comprise markets,

firms and other institutions which facilitate and constrain economic interactions.

Institutions have four key dimensions, which receive differing degrees of emphasis in the

various strands of institutionalism: a behavioural or regulative dimension (providing the

norms of action which are externally enforced), a cognitive dimension (beliefs and

understandings of cause and effect relations), a normative dimension (providing norms of

appropriateness and legitimacy), and, some would add, a resource dimension (distributing

resources and regulating access and agendas of decision making).24

The core notion is that ‘institutions matter’ to individual and social action and interaction

because they provide the structure in which the action and interaction occurs. They

provide shared conceptions of reality, meaning systems and collective understandings

that guide decision making and which individuals take for granted. Actors articulate and

define their policy problems and solutions by using institutionalised scripts, cues and

routines that constitute their cognitive frameworks and empower them to act, but on

which they do not necessarily reflect.25 Decisions are made to pursue goals, but often the

23Ikenberry, G. (1988), ‘Conclusion: An Institutional Approach to Foreign Economic Policy’ in G. Ikenberry, D.A. Lake and M. Mastanduno (eds), The State and American Foreign Economic Policy, Ithaca: Cornell University Press; Jepperson, R.L. (1991), ‘Institutions, Institutional Effects and Institutionalism’ in W.W. Powell and P.J. DiMaggio, The New Institutionalism in Organisational Analysis, Chicago: University of Chicago Press; W. R. Scott op.cit; DiMaggio, P.J., and W.W. Powell (1991), ‘Introduction’ in W.W. Powell and P.J. DiMaggio (eds), The New Institutionalism in Organizational Analysis, Chicago: Chicago University Press.24 March, J. and J. Olsen,(1984), ‘The New Institutionalism: Organisational Factors in Political Life’, American Political Science Review, 78: 734-749; Hall, P. and D. Soskice (2001b), ‘An Introduction to Varieties of Capitalism’ in P. Hall and D. Soskice (eds) (2001a), Varieties of Capitalism: The Institutional Foundations of Comparative Advantage, Oxford: Oxford University Press; W.R. Scott, op.cit. 25 DiMaggio and Powell, op.cit; W.R. Scott, op.cit; Meyer, J.W., J. Boli and G.M. Thomas (1994), Ontology and Rationalization in the Western Cultural Account; in W.R. Scott, J.W. Meyer, and Associates, Institutional Environments and Organizations: Structural Complexity and Individualism, Thousand Oaks,

reaffirmation of processes and rituals and the communication of symbols and legitimacy

is equally if not more important.26 Finally, institutions have both stabilising and

facilitating effects, for example they explain how collective action problems are

overcome; the stability of political decision making; and why, in a situation of multiple

Pareto-optimal equilibria, one policy option is chosen over another.

In institutional theory, legitimacy is central to an organisation’s survival and

development. Legitimacy means social credibility and acceptability: ‘a generalized

perception or assumption that the actions of an entity are desirable, proper, or appropriate

within some socially constructed system of norms, values, beliefs, and definitions.’27

This understanding of legitimacy has a clear link with that set out above: that legitimacy

is a perceived right to govern.

The notion of accountability posited above, as relational, can also be developed from an

institutionalist base. The argument presented here is that accountability is not an abstract,

technical process. This conception of accountability runs counter to the ‘invisible hand’

of cybernetics and collibration, or other mechanistic portrayals of accountability.

Accountability is rather relational, and dialogic. Here again it is important to emphasise

its distinctiveness from control or other institutions or events which affect behaviour.

Firms may respond to consumer demand by developing ethical products, but changes in

their production processes or sourcing policies to respond to consumer demand does not

necessarily mean they are ‘accountable’ to consumers; rather they are influenced by

them. Being ‘accountable’ would mean that firms create structures through which they

can communicate with consumers, or more realistically groups of consumers or those

purporting to act on their behalf, and through which consumers can respond in such a

way that will have consequences for the firm’s behaviour.

Calif., Sage.26 Meyer, J.W. and B. Rowan (1977), ‘Institutionalized Organizations: Formal Structure as Myth and Ceremony’, American Journal of Sociology, 83, 340-363.27 W.R. Scott, op. cit.

Institutionalism lacks a strong communicative dimension, however. Here, discourse

theory can be used to complement institutionalist analysis in a number of ways;

institutionalism in turn provides analytical structures which explain how discourse can

affect action more satisfactorily than discourse theory can itself.28 A discourse theory of

regulation argues that discourse constitutes regulation in that it builds understandings and

definitions of problems (eg ‘market failure’, ‘risk’, ‘accountability gap’) and acceptable

and appropriate solutions (‘meta-regulation’, ‘precautionary principle’, ‘audit’,

‘participation’,), it builds operational categories (eg ‘compliance’, ‘transparency’), and

produces the identities of and relations between those involved in the process. It is

functional in that it is designed to achieve certain ends (eg the strategic use of rule design;

the deployment of skills of argumentation and rhetoric by all involved at every stage). It

is co-ordinating in that it produces shared meanings as to regulatory norms and social

practices which then form the basis for action (eg the formation of regulatory interpretive

communities, or of accountability relationships).

Discourse theory therefore suggests that different legitimacy claims, and associated

accountability relationships will be grounded and expressed in different discourses; that

these discourses will build different understandings and definitions of the ‘accountability’

problem and appropriate solutions, such as audit, performance evaluation, participation,

and operational categories, such as compliance, so as to produce different forms of

accountability relationship. Most importantly for this discussion, discourse theory

emphasises the role of power in regulation: discursive practices, events and texts arise out

of and are shaped by power and ideology and struggles over them, suggesting that which

legitimacy claims an organization will respond to will be a reflection of the constellation

of power relationships in which it is situated.

Competing claims of accountability and legitimacy

28 See J. Black,’Regulatory Conversations’ (2002) Jnl Law and Society for elaboration.

Thus far it has been argued that legitimacy is socially constructed through institutions,

that accountability is a communicative relationship entailing responsiveness, and that

different legitimacy claims are expressed in different legitimacy discourses. This

analysis is to an extent consistent with the ‘principal-agent’ conception of accountability

relationships familiar from some political science,29 in that relationships of authority are

argued to underly accountability relationships, although it prefers the less dichotomous

model of organisation and environment.

But where does this leave the familiar debates on the ‘types’, ‘forms’ or ‘sources’ of

accountability and legitimacy? Categorisations of accountability are as numerous as its

definitions. The familiar ones include: legal, financial, managerial, ethical, technical,

democractic and procedural. Legitimacy criteria are just as well numerated and have

been argued to include: legal mandate, due process, efficency, effectiveness and

expertise,30 to which may be added representativeness and / or democratic manadate, and

conceptions of justice.

Categorisation exercises can be helpful; they can facilitate analysis and thus aid

understanding. However what the analysis above suggests is that categorisation in itself

does not answer the question: is this body accountable or legitimate, because those

questions are inherently relational. They cannot be answered in the abstract. They have

to be met with response: accountable or legitimate to whom? Discussions on

categorisation which link types of accountability with different values go further in this

respect, as do those which recognise that different forms of accountability conflict. But

again the essentially constructed nature of accountability, and indeed legitimacy, is

glossed over.

Nonetheless, these different typologies do draw attention to the differentiated, and indeed

contested, nature of both accountability and legitimacy. Organisational environments are

not homogenous, and what is considered to be ‘desirable, proper or appropriate’ varies 29 Eg R. Keohane and J. Nye, Power and Interdependence (3rd ed, New York: Addison-Wesley Longman, 2001).30 For discussion see R. Baldwin and J.C. McCrudden, Regulation and Public Law (London: Weidenfeld and Nicholson, 1987); and R. Baldwin and M. Cave, Understanding Regulation (Oxford: OUP, 1999).

considerably across them. The implication of this analysis is that what constitutes

legitimacy for one part of an organisation’s environment, or what I characterize here as

different legitimacy communities, will differ from that which constitutes legitimacy for

another. Moreover, different legitimacy communities may well be seeking different

accountability relationships, directly or indirectly, with the relevant organization.

It is suggested that there are four broad types of legitimacy claims31 and associated

accountability relationships: legality / regulative, justice, performance and

representation.32 Legitimacy based on legality refers here very broadly to conformance

with written norms (thus embracing so-called ‘soft law’ or non-legal, generalised written

norms), and actions in conformity with a procedural form of the ‘rule of law’, including

legal values of procedural justice. Its application beyond law is signalled by the

alternative term ‘regulative’. Legitimacy claims rooted in justice refer to the normative

values which the organization is pursuing, including the conception of justice

(republican, Rawlsian, utilitarian, for example, or the various religious conceptions of

‘truth’ or ‘right’ ). Performance refers to the manner of functioning of the organization:

its conformity to ethical, managerial, financial, technical norms. Representation refers to

the conformance of the organization to different governance models: representative

democracy, deliberative democracy and so on. Each of these broadly corresponds to the

challenges outlined above: systemic, functional, normative and democractic.

All of these merit considerable elaboration in themselves; however, the point that I want

to make here is that these legitimacy claims are both contested and contestable, not only

between the different groups, but within them. Indeed, there may be more interlinkages

between the different groups than there are within them: between procedural justice as

understood by law and the pursuit of conceptions of substantive justice, for example, or

between deliberative democracy and particular conceptions of justice; or between legality

and different forms of performance (eg ethical performance in legal organizations). In

contrast, the differences between managerial and ethical claims to legitimacy, and

31 Legitimacy claims refer to claims made by the organization for legitimacy and the bases of credibility or acceptability that an organization has in different parts of its environment.32 See Dubnick for a similar breakdown.

associated accountability relationships, can compete to a considerable degree – indeed

exploring the tension between them is a fertile ground of research in public

administration, and there have been many studies on the impact of managerial modes of

accountability on ethical performance in professional organizations, notably in the fields

of health and education.

Moreover, different legitimacy discourses will build, and be reinforced by, different types

of accountability relationships. Those of legality require, for example, conformance to

written norms, an impartial forum where disputes over the application of those norms is

determined, and rules to be clear, stable and prospective and to be applied impartially and

in accordance with norms of fair procedure. There are in turn differences over just what

conception of fair procedure, or what constitutes impartiality and so on. Accountability

relationships which stem from discourses of functionality include, for example, those of

audit, cost-benefit analysis, performance evaluation. Legitimacy claims related to

representativeness require accountability relationships based on different conceptions of

representative, participative or deliberate democracy, for example. Accountability

relationships associated with legitimacy claims related to justice are built around

competing conceptions of rights or ‘the good’. Each of these broad groups of legitimacy

claims has internal fractures and divisions, resulting in ever more complex demands for

different types of accountability relationships.33

Organisational responses to competing claims

It is not necessary for legitimacy claims to be interlinked or mutually supporting for it to

be possible for them to be met by an organization simultaneously. They simply have to

be compatible. It may be, for example, that for a firm the ‘win-win’ solution of

sustainable development and profitability can be attained, for example; or that a

33 In terms of how this maps on to the four types of control mechanism outlined by Hood et al: hierarchy, mutuality, competition and random selection – accountability relationships stemming out of any of the legitimacy claims can take these forms. These four types of control mechanism can be seen as cutting across the different accountability relationships and legitimacy claims. See C. Hood et al, Regulation Inside Government (OUP).

regulatory organization can operate in accordance with accepted modes of procedural

justice whilst also operating in a financially responsible manner. The development of

management models such as balanced scorecards, used by a number of UK government

bodies, are just such attempts to structure, or at least provide a reporting framework, for

the organisation’s responsiveness to different aspects of its environment.

However, the demands of legitimacy communities may well be directly opposed, such as

to satisfy one will necessarily lead to dissatisfaction of the other. Representativeness is a

good example. Non-state regulators face this problem to a significant degree.

Frequently, in order to satisfy the legitimacy claims of those they are regulating, their

main decision making bodies need to be comprised solely or mainly of representatives of

those regulatees; however to be legitimate to a wider section of civil society, those

decision making bodies need to be solely or mainly composed of representatives from

other sections of civil society. The same debate occurs in the context of risk regulation:

to be legitimate to scientists, regulators have to be composed of scientific experts or at

least be governed by them in their decisions; however to be legitimate to other sectors of

society, regulators have to include a wider proportion of society in their decisions – the

familiar debate on lay versus expert models of decision maker. There are other conflicts:

between those demanding procedural justice and those demanding maximum speed and

efficiency in decision making. Even within models of administrative justice there are

conflicts, as Mashaw’s familiar typology of bureaucratic, moralistic and professional

models of administrative justice demonstrates.34

Organisations then face a legitimacy dilemma: what they need to do to be accepted by

one part of their environment is contrary to how they need to respond to another part.

Forming one set of accountability relationships can preclude forming others; it simply is

not possible to have complete legitimacy from all aspects of its environment.

Even if the conflict between legitimacy communities does not lead to a dilemma, it can

have a deleterious effect on the organization as it seeks to respond to the multiple

34 J.L. Mashaw, Bureaucratic Justice: Managing Social Security Disability (New Haven, 1983).

legitimacy and accountability demands being made on it: in Koppell’s evocative phrase,

the organization may suffer ‘multiple accountability disorder’.35 In other words, its

attempts to respond to the multiple demands may diminish its chances of survival.

Implications for the accountability and legitimacy of non-state regulators

What are the implications of this analysis for our understanding of the accountability and

legitimacy of non-state regulators? It is suggested that two main implications flow which

have not yet received the research attention that they merit.

The first implication relates to the debate on the design of accountability mechanisms.

These debates assume that accountability is a technical issue; that what is necessary is to

improve the design, enhance the mechanisms, structure the springs, in such a way that

accountability will necessarily follow. But accountability is not a technical exercise; it is

a relational one.36 Different accountability mechanisms are different accountability

relationships. Those relationships are not necessarily substitutable one for the other such

that if one fails another can take its place. Substitutability assumes homogeneity in

legitimacy claims within the organisation’s environment. But as discussed above, those

claims are heterogenous, and moreover that incompatibilities can be greater within the

broad forms of legitimacy claims than between them.

Far more could be said on that issue. However it is the second set of implications on

which I want mainly to focus. That is what the above analysis implies for the

organisation which is the subject of the accountability and legitimacy claims. How does

it respond to these competing claims? Here there is very little empirical research done

on non-state regulators. There is more in the field of CSR, and to an extent in regulation,

where there is a considerable body of research investigating questions such as why do

35 J. Koppell, ‘Pathologies of Accountability’: ICANN and the Challenge of ‘Multiple Accountability Disorder’ (2005) 65(1) Public Administration Review 94.36 See also S. Lister, ‘NGO Legitimacy: Technical Issue or Social Construct? (2003) 23(2) Critique of Anthropology 175.

firms develop environmental policies which extend beyond regulatory requirements,37

why do they comply with non-legal codes of conduct relating to corporate governance,

and so on. There has also been research on the narrower question of the impact of one

type of accountability, new public management reforms, on parts of the civil service and

public service organisations in health and education. There is some work which focuses

on INGOs, though for the most part this has been of the ‘powering’ type – offering

blueprints for accountability and legitimacy, blueprints which themselves are usually

rooted in particular legitimacy claims (managerial performance, participative

democracy). However there has been very little sustained research on how non-state

regulators respond to these competing accountability and legitimacy claims, be they of

the ‘agency’ type, or be they organisations which are seeking to apply common norms

but without a central organisational structure to the regime (as in the case of the Equator

Principles noted above).38

Here it is suggested that Oliver’s typology of organisational responses could be a useful

starting point for such an empirical inquiry. In a synthesis of resource theory of

organisations and institutional theory,39 Oliver identifies five types of strategic response

by organisations to institutional processes.40

Table 1: Strategic Responses41

Strategies Tactics Examples

Habit Following inevitable, taken for granted norms

37 On the latter see R. Kagan, N. Gunningham and D. Thornton, ‘Explaining Corporate Environmental Performance: How Does Regulation Matter?’ (2003) 37 Law and Society Review 51. 38 It should be noted that there is also little research on how organisations respond to competing regulatory norms, both legal and non-legal; for an exception see Haines and ?. Most research is done on a domain-specific basis: ‘how did Organisation X respond to set of regulatory norms Y?’ where ‘Y’ is environmental rules, competition requirements and so on. This is also true for studies with transnational regulation, both treaty-based and ‘soft law’ provisions, see eg D. Shelton (ed), Commitment and Compliance (OUP, 2000).39 And assuming a particular model of action: Giddens’ structuration (in which an organisation or individual has agency but that agency is structured by their institutional environment).40 C. Oliver, ‘Strategic Responses to Institutional Processes’ (1991) 16(1) Academy of Management Review 145.41 Ibid, p. 152.

Acquiescence Imitate Mimicking institutional models

Comply Obeying rules and accepting norms

Balance Balancing the expectations of multiple constituents

Compromise Pacify Placating and accommodating institutional elements

Bargain Negotiating with institutional stakeholders

Conceal Disguising nonconformity

Avoid Buffer Loosening institutional attachments

Escape Changing goals, activities or domains

Dismiss Ignoring explicit norms and values

Defy Challenge Contesting rules and requirements

Attack Assaulting the sources of institutional pressure

Co-opt Importing influential constituents

Manipulate Influence Shaping values and criteria

Control Dominating institutional constituents and processes

Acquiescence is accession to institutional pressures. Such acquiescence may be ‘blind’ in

the sense that it is an unconscious adherence to taken for granted rules; or it may be more

strategic – compliance with regulatory norms, for example, in order to gain credibility

and thus reduce negative assessments of its conduct, product or services. Compromise is

an attempt to balance, pacify or bargain with external constituents, where demands are

competing, and / or where the organisation considers unqualified conformity unpalatable

or unworkable. Avoidance is the attempt to preclude the need to conform, by concealing

non-conformity, buffering itself from institutional rules or pressures by decoupling its

technical activities from external contact, or escaping from the domain in which pressure

is being exerted: ‘exit’ in Hirschmann’s terms. Defiance is a more active form of

resistance to institutional pressures, and can consist of dismissing or ignoring institutional

rules or values; challenging them; or more directly attacking them. Finally, organisations

may manipulate the environment by actively changing, or seeking to change or exert

power over the content of the expectations themselves or the sources that seek to express

or enforce them.42 This may be through co-opting the source of the pressure, attempts to

manipulate the terms of debate,43 or using controlling tactics: specific attempts to exert

power or dominance over the external constituents that are applying pressure.

What determines which approach an organisation will adopt is they key research

question. Response will be a function of both capacity and willingness to conform to the

institutional environment, or at rather different parts of it. Oliver suggests that the scope

of an organisation’s willingness to respond is bounded by organisational scepticism,

political self-interest and organizational control. The scope of an organisation’s capacity

to respond is bounded by organisational capacity, internal conflict and awareness. She

suggests that organisational responses to institutional pressures to conformity will depend

on five sets of categories, each with two dimensions:

the nature of the pressures exerted / what the organisation gains from

acquiescence (cause)

o social fitness

o economic gain;

who is exerting them (constituents)

o the multiplicity of claimants and

o the organisation’s dependence on them;

how consistent are those demands with the organisation’s own goals (content)

o whether they are congruent with the organisation’s goals and

o whether they constrain the organisation’s discretion or not;

the means by which they are imposed (control)

o by legal or non-legal means; and

o whether the norms, values and practices are diffused or not throughout the

environment42 Oliver, op.cit. pp 152-159.43 Oliver does not note this, but discourse theory would suggest that this could be through deploying rhetorical tactics or otherwise seeking to change the discursive framing of an issue.

the nature of the environmental context in which they occur (context)

o whether it is characterised by uncertainty, and

o the degree of interconnectedness or degree of inter-organisational relations

within the organisational field.44

She then draws on sociological institutional theory, supplemented by organisational

resource theory, to develop hypotheses on choice of strategy based on variation in the ten

dimensions of these five categories.

Table 2: Oliver’s Predictions of Organisational Responses

Strategic

Response /

Predictive

Factor

Acquiesce Compromise Avoid Defy Manipulate

Cause

Legitimacy High Low Low Low Low

Efficiency High Low Low Low Low

Constituents

Multiplicity Low High High High High

Dependence High High Moderate Low Low

Content

Consistency High Moderate Moderate Low Low

Constraint Low Moderate High High High

Control

Coercion High Moderate Moderate Low Low

Diffusion High High Moderate Low Low

44 Ibid, p.159.

Context

Uncertainty High High High Low Low

Interconnectedness High High Moderate Low Low

The key sets of hypotheses for our purposes is how will an organisation respond in a

situation of uncertainty, in which uncertainty is increased by multiple, mutually

incompatible or inconsistent demands,45 for that is the situation faced by many

organisations, not least non-state regulators. Here the hypothesis suggests that responses

will range from compromise to manipulation, depending on the degree of dependence,

consistency, constraint, coercion, diffusion, uncertainty or interconnectedness in the

organisation’s environment. However, there is no set of hypothesis for the situation

where legitimacy gains are perceived to be high in such a situation, presumably because

overall legitimacy gains cannot be high in a situation of multiple and competing

demands. However, it may still be the case that legitimacy gains may potentially be

made to a number of different legitimacy communities (ie is it always ‘high’); but as it

stands Oliver’s hypotheses do not extend to situations where although legitimacy gains

might be high with respect to particular legitimacy communities, the organisation does

not acquiesce. Arguably the hypothesis here would be that in a situation of multiple

legitimacy claims, an organisation will not acquiesce even if legitimacy gains are high in

situations where there is a low dependency of the organisation on that legitimacy

community; the normative content of the claims are congruent only to a moderate or low

degree; the constraints on discretion would be moderate or high; coercion is low;

diffusion is low; uncertainty is high and interconnectedness is low.

Clearly therefore there is scope for elaboration and refinement of Oliver’s model, but it is

suggested that it could found the basis of fruitful empirical research into how non-state

regulators in fact respond to demands for legitimacy and accountability.

45 Incompatible here means the demands are by nature incompatible; inconsistent means that they are potentially compatible but are differently demanded over time.

Still lacking from this analysis is any real conceptualisation of the role of power in

accountability and legitimacy relationships – something which is often underplayed in

sociological institutionalism.46 However, it is suggested that the introduction of discourse

theory into the analysis redresses this omission, and in turn has implications for those

who are seeking to make the organisation respond to their legitimacy claim. It suggests

that absent any other resource, the legitimacy community (or its constituent) will have to

rely on discourse; in other words it will have to translate its demands into a discourse that

the organisation already recognises in order to build relationships of accountability with

the organisation that it seeks.47

Summary

This paper did not set out to provide an answer directly to the extraordinarily difficult

questions: how to make actors in decentred regulatory regimes, in particular, non-state

regulators, accountable. It seeks to clear a path to answering it, however, by proposing

that we understand better how such organisations themselves respond to the various

legitimacy demands which are made upon them. Not only is such understanding of

intrinsic interest, it should form the basis for forging answers to that initial, challenging

question.

In order to begin to create that path, some conceptual groundwork had to be done. The

paper therefore examined the relationship between accountability and legitimacy, and

suggested that both are relational concepts which are socially and discursively

constituted. This in turn provides the basis for the second proposition, which is that there

are different legitimacy communities, comprised of actors within and outside the

regulatory regime who have different perceptions as to the relevance and validity of

different legitimacy claims with respect to different regulatory actors. This in turn led to

a third proposition, which is that different legitimacy claims, and associated discourses,

are not always compatible but may compete. Fourthly, that although organisations can

46 For emphasis elsewhere see Keohane and Nye, op. cit.47 Eg M. Hajer, The Politics of Environmental Discourse (1995); B. Morgan, Social Citizenship in the Shadow of Competition (Ashgate, 2003).

often participate in a number of different legitimacy discourses simultaneously, and thus

satisfy a range of different legitimacy communities, this can not only have a deleterious

affect on the organisation (which may suffer ‘multiple accountability disorder’) the

differences between communities may be such that organisations can face a legitimacy

dilemma: that actions that they need to take to render them legitimate for one legitimacy

community are in direct opposition from those they need to adopt to satisfy another.

These propositions already find support in much of the current debate. However, the

paper then sought to explore two main consequences of these propositions for th debate

on regulatory accountability and legitimacy. Firstly, that different accountability or

legitimacy mechanisms (or rather relationships) are not necessarily substitutable, as not

all will satisfy every legitimacy community. It is not therefore always possible maintain

legitimacy by replacing one with another when the first one fails, unless the replacement

is recognised by that legitimacy community. Secondly, that how organisations respond

to competing legitimacy demands is structured by the particular institutional context in

which the regulatory regime, and the individual organisation, operates. Regulators are

not ciphers. They can be active participants in the debate on their own accountability and

legitimacy, not just passive recipients. They may exhibit the same strategies of

avoidance, defiance, manipulation, compromise or acquiescence in response to pressures

for their accountability and legitimacy as any organisation does in response to any norms

which others seek to impose on them. ‘How to’ proposals for accountability which

ignore these different organisational response strategies are as weak as any proposal for a

regulatory strategy which ignores how its prescriptions will be received.

Finally, the paper proposes a further theoretical step and, pursuing the interplay of

institutionalism with discourse theory, suggests that in acquiescing in, or at least reaching

a compromise with, some legitimacy communities, the organisation indicates that it is

prepared to recognise those, whilst refusing to recognise others. This has implications for

those who are seeking to get the organisation to respond to their accountability and

legitimacy demands. Absent other key resources, the paper suggests that those

legitimacy communities whose demands have not been recognised will have to translate

them into the discourses of those that have been recognised if they are to build the

relationships of accountability with the organisation that they seek. Together, it is

suggested that these propositions could form the basis of detailed empirical research into

how non-state regulators respond to competing demands for legitimacy and

accountability, particularly in situations of uncertainty.